Dissecting the Role of Mast Cells in Mouse Models for Human

Collectively, my results indicate that, although in vitro A20 deficiency causes dramatically enhanced NF-κB activation in response to various physiologically relevant stimuli, under steady-state conditions mast cell-specific A20 deficiency does not cause overt spontaneous inflammation but induces a pre-activated or poised state.

7.1.3. Dissecting the Role of Mast Cells in Mouse Models for Human

In this context, it is interesting to note that various studies have suggested a pathogenic role for IL-33 in both asthma and rheumatoid arthritis (Ohno et al., 2012;

Palmer and Gabay, 2011). Additionally, it was proposed that during the course of arthritis, IL-33 could promote joint inflammation at least in part by the activation of mast cells (Xu et al., 2008). This is in line with my findings that loss of A20 in mast cells exacerbates IL-33-induced NF-κB activation and pro-inflammatory cytokine production (Publication I). Hence, augmented responses to IL-33 might lead to locally restricted inflammation during arthritis pathology. In this scenario, enhanced IL-33-dependent pro-inflammatory cytokine production by A20-deficient mast cells might in turn cause greater IL-33 expression in synovial fibroblasts. Its subsequent release could signal back to mast cells leading to paracrine auto-amplification and a detrimental feed-forward loop. This scenario is supported by the fact that mast cell-specific MyD88 deficiency ameliorates arthritis onset as well as disease symptoms (Publication I).

In addition to arthritis, IL-33-dependent paracrine feed-forward loops might also play a central role in various other inflammatory diseases that have an IL-33 component and are associated with polymorphism in the A20 gene locus, such as psoriasis, systemic sclerosis and inflammatory bowel disease (Ma and Malynn, 2012;

Palmer and Gabay, 2011; Vereecke et al., 2009). Interestingly, mast cells have also been implicated in these pathologies (Chichlowski et al., 2010; Eklund, 2007;

Harvima et al., 2008). Hence, dissecting mast cell-specific A20-deficieny in relevant mouse models might yield important insights into the etiology and progression of these diseases.

My results are in apparent discrepancy to a study showing that the loss of mast cells in Kit^W-sh/W-sh mice did not significantly alter CIA disease outcome (Pitman et al., 2011). However, studying the function of mast cells by their absence in Kit^W-sh/W-sh mice bears the inherent caveat that other cell types could functionally compensate for their loss. This has already been demonstrated during the early innate response to haptens, in which neutrophils take over the function of mast cells in the context of the prominent myelodysplasia seen in Kit^W-sh/W-sh mice (Dudeck et al., 2011). Hence, the depletion of neutrophils would clarify their contribution in Kit^W-sh/W-sh mice during CIA.

In contrast to airway and joint inflammation, A20-deficient mast cells did not worsen EAE symptoms (Publication I), possibly because initiation and progression of

completely exclude a role for mast cells in this model. Firstly, deletion of A20 is limited to CTMCs while MMCs remain unaffected and, secondly, A20 deficiency selectively influences pro-inflammatory NF-κB-driven responses without affecting degranulation. Hence, general activation of mucosal in addition to connective tissue mast cells might have different effects than seen in Mcpt5Cre A20^F/F mice.

Nonetheless, my results clearly show that enhancing the pro-inflammatory reactions of CTMCs has no influence on MOG peptide-induced central nervous system inflammation. These results are in contrast to the apparently well-documented role of mast cells in EAE pathology (Sayed et al., 2008; Secor et al., 2000). However, this notion is primarily based on a single genetic tool, namely the use of mast cell-deficient Kit^W/Wv mice. In recent years the use of mast-cell-deficient Kit^W-sh/W-sh mice has become more and more popular as this strain shows milder Kit-dependent developmental abnormalities than Kit^W/Wv mice (Grimbaldeston et al., 2005).

Comparing mast cell-deficient Kit^W/Wv and Kit^W-sh/W-sh mice yielded conflicting data regarding their contribution to autoimmune pathologies. Surprisingly, Kit^W-sh/W-sh mice showed even earlier onset and developed exacerbated EAE symptoms (Li et al., 2011; Piconese et al., 2011). Furthermore, depending on the immunization protocol, EAE disease outcome can be modulated in Kit^W/Wv mice from protective to even exacerbating (Piconese et al., 2011) and two recent reports failed to confirm that Kit^W/Wv mice are protected from EAE (Bennett et al., 2009; Feyerabend et al., 2011). It is worth noting that all of these results were obtained in the absence of mast cells and fail to clarify whether or not mast cells participate in EAE disease pathology.

Moreover, my results using a gain-of-function approach by mast cell-specific ablation of A20 (Publication I) and novel Kit-independent mast cell-deficient mice have failed to corroborate a contribution of mast cells to EAE (Feyerabend et al., 2011).

Collectively, these results do not support a major function of mast cells in the pathogenesis of EAE.

In addition to playing central pathological roles in autoimmune diseases, many reports assign to mast cells various additional effector and regulatory functions in immune responses against viral (Aoki et al., 2013), bacterial (Echtenacher et al., 1996; Malaviya et al., 1996), and parasitic infections (Ha et al., 1983), in cardiovascular disorders (Sun et al., 2007), obesity and diabetes (Liu et al., 2009), in malignant diseases and angiogenesis (Coussens et al., 1999), in contact hypersensitivity in the skin (Grimbaldeston et al., 2007) and the recognition of cell injury (Enoksson et al., 2011; 2013). However, most in vivo evidence for these mast

cell functions is also based on Kit mutant mast cell-deficient mice. Similar to discrepant results on the role of mast cells in autoimmune diseases, studies using novel Kit-independent strains failed to confirm some of the initial results regarding contact hypersensitivity (Dudeck et al., 2011; Otsuka et al., 2011), wound healing (Antsiferova et al., 2013) and skin carcinogenesis (Antsiferova et al., 2013). These fundamental discrepancies indicate that using Kit mutant mice and the mast cell reconstitution approaches can generate misleading results and might lead to false interpretations. Therefore, it will be essential to readdress central mast cell findings using novel genetic mouse models. My analysis on the consequences of A20-deficiency in mast cells yielded specific results with regard to different disease models, specifically, exacerbation of innate forms of asthma and CIA in contrast to no effect on EAE. This shows that the Mcpt5Cre A20^F/F strain will be a valuable and informative gain-of-function tool to complement loss-of-function studies. This should be especially true under conditions where mast cells are presumably activated through their IL-33R or TLRs such as tumor development, the initiation of Th2 responses, and the detection of tissue damage or microbial infections.

In summary, my data demonstrate that loss of A20 specifically in mast cells provides a novel genetic model system to study their pro-inflammatory properties in a gain-of-function approach. These hyperactive mast cells exacerbated allergic airway and skin responses, as well as autoimmune joint inflammation pointing to an important contribution of mast cells in these pathologies.

Inducible Conditional Gene Targeting in Mast Cells 7.2.

Allows Studying their Differentiation and Cellular Maintenance

When I started my thesis, mast cell-specific conditional gene targeting was in its early stages and the functions of specific genes were studied by the reconstitution of Kit mutant mice with cultured mast cells derived from mutant or transgenic bone marrow (Tsai et al., 2005). However, depending on the route of administration, reconstitution may not reach physiological mast cell levels or their number might sometimes even be higher than normal (Grimbaldeston et al., 2005; Nakano et al., 1985). In addition,

2012). Therefore, using this mast cell reconstitution approach does not allow faithful investigation of the function of specific genes during bona fide mast cell development, homeostasis and function. Recent advances in gene targeting and transgenesis led to the generation of novel mouse strains that constitutively express the Cre recombinase and, thus, allow mast-cell-specific conditional gene targeting to varying degrees of efficiency and specificity (Feyerabend et al., 2011; Furumoto et al., 2011;

Müsch et al., 2008; Scholten et al., 2008). However, if the loss of specific genes impedes or affects mast cell development, temporal control of Cre activity is required to delete genes in mature mast cells. Furthermore, acute gene deletion is essential for the validation of potential drug targets and the avoidance of molecular and cellular compensation mechanism. Similarly, acute gene modification is required for in vivo cell-fate tracking, a powerful genetic approach to study cellular differentiation.

The group of Dr. Dieter Saur has generated a knock-in mouse strain expressing a tamoxifen-inducible version of the Cre recombinase from within the endogenous c-Kit locus (Klein et al., 2013). Since Cre efficiency is relative to its expression strength (Araki et al., 1997; Schmidt-Supprian et al., 2007) and mast cells within the hematopoietic system exhibit very high c-Kit expression (Publication III), I reasoned that this mouse strain should allow efficient recombination of conditional (loxP-flanked) alleles in mast cells. Indeed, the use of a fluorescent Cre recombinase activity reporter (R26-Stop^FYFP) revealed high mast cell specificity and efficiency and gave important insights into mast cell development.

Mast cells are thought to share a common progenitor with basophils in the spleen (basophil and mast cell progenitor, BMCP) (Arinobu et al., 2005). However, although more than 90% of mature dermal and peritoneal mast cells, and nearly 16%

of BMCPs in tamoxifen-fed Kit^CreERT2/+R26-Stop^FYFP animals expressed YFP, recombination in basophils was minimal (Publication III). This indicates that, under physiological conditions in the steady-state, YFP-positive BMCPs do not significantly differentiate into basophils. My results hence challenge the validity of BMCPs as authentic bipotential basophil and mast cell progenitors. This is in line with two recent publications demonstrating a greater tendency of BMCPs to develop into mast cells in vitro (Mukai et al., 2012; Qi et al., 2013). Further pulse-labeling experiments demonstrated that 8 weeks after cessations of tamoxifen, still more then 90% of mast cells expressed YFP (Publication III). These results indicate a very slow turnover of mature mast cells and regeneration from unaffected progenitors. To the best of my knowledge, this is the first direct genetic evidence that mature mast cells are indeed

very long-lived. Previous attempts to study mast cell turnover by radioactive thymidine labeling provided half-life estimates ranging from days up to years (Kiernan, 1979; Walker, 1961). Although the exact reasons for mast cell longevity are presently unclear, it has been speculated that mast cells might be able to modulate the content of their granules in response to certain stimuli, a process that could be considered to be a form of immunological memory (Abraham and St John, 2010).

Moreover, I found that 4 weeks after ablation of mast cells by DTA expression (Kit^CreERT2/+R26-GFPStop^FDTA), very few mast cells had regenerated from unaffected progenitors (Publication III). These results are in full agreement with a recent study by Dudeck and colleagues who demonstrated little mast cell recurrence within 3 weeks after their selective ablation using a different genetic system (Dudeck et al., 2011). Both results corroborate the long-standing concept that, under steady-state conditions, mature connective tissue mast cells are rarely replaced by progenitors from the bone marrow (Kitamura et al., 1977). This is reminiscent of microglia in the central nervous system, Langerhans cells in the epidermis, Kupffer cells in the liver and macrophages in the serosal cavities (Gomez Perdiguero and Geissmann, 2013).

These cells all originate from the yolk sac and their precursors migrate out into the periphery predominantly during embryonic development (Gomez Perdiguero and Geissmann, 2013). Similarly, mast cell colonies can be generated from yolk sac cells (Sonoda et al., 1983) and mast cell precursors peak in fetal blood around day 15 of gestation and subsequently seed the fetal skin (Hayashi et al., 1985; Rodewald et al., 1996). However, it is presently unknown to what extent mast cells present in the adult are derived from this embryonic wave. Administration of 4-hydroxytamoxifen to pregnant R26-Stop^FYFP females crossed to Kit^CreERT2/+ males might allow mast cells and their c-Kit⁺ precursors to be pulse-labelled during embryogenesis and, hence, might give important insights into the developmental origin of mature mast cells.

In summary, the novel Kit^CreERT2 mouse strain will be very useful to define the role of specific genes during physiological and pathological mast cell functions. The inducible nature of the Cre recombinase will furthermore facilitate studies of mast cell development, homeostasis and turnover.